We propose to study changes in household and families in Latin America during the period 1970-1985. We suggest the application of a framework that considers these changes as the result of demographic factors (fertility, mortality, nuptiality, and migration) and a component reflecting propensities, preferences, or choices. These two components mediate the impact of socio-economic transformations on household arrangements. The study is designed to falsify hypotheses drawn from theories of family, kin relations and household change that have been primarily formulated to explain secular transformations in Western Europe and North America. With a combination of censuses and household surveys for the period 1970-1985 in fourteen Latin American countries we will examine several dimensions of household and families: household and family extension; prevalence and types of female-headed households; the household arrangements of older people, the household composition of children; the timing of leaving home by young adults; being "unrelated" to the head of the household; and solitary living. We propose to use a mixture of approaches including multivariate analysis, macro simulation (FAMY) and micro simulation (SOCSIM and MOMSIM). Our goal is to utilize these approaches to test theories while properly controlling for the effects of demographic factors. In particular, we suggest alternative ways of decomposing household changes into demographic and non-demographic components. Although micro and macro simulation models are not designed to provide an unambiguous solution to the difficult problem of identifying unobserved preferences from realized arrangements, their judicious use and application leads to a range of estimates useful for hypotheses verification. Indeed, we show that the combination of these analytic strategies can overcome at least partially the thorny problem posed by censuses and surveys that are silent on the existence and characteristics of most kin who do not co-reside.